From Dr. Roy Spencer’s Blog

January 14th, 2020 by Roy W. Spencer, Ph. D.

In 2017, Christy & McNider published a study where they estimated and removed the volcanic effects from our UAH lower tropospheric (LT) temperature record, finding that 38% of the post-1979 warming trend was due to volcanic cooling early in the record.

Yesterday in my blog post I showed results from a 1D 2-layer forcing-feedback ocean model of global-average SSTs and deep-ocean temperature variations up through 2019. The model is forced with (1) the RCP6 radiative forcings scenario (mostly increasing anthropogenic greenhouse gases and aerosols and volcanoes) and (2) the observed history of El Nino and La Nina activity as expressed in the Multivariate ENSO Index (MEI) dataset. The model was optimized with adjustable parameters, with two of the requirements being model agreement with the HadSST global temperature trend during 1979-2019, and with deep-ocean (0-2000m) warming since 1990.

Since the period since 1979 is of such interest, I re-ran the model with the RCP6 volcanic aerosol forcing estimates removed. The results are shown in Fig. 1.

Fig. 1. 1D model simulation of global (60N-60S) average sea surface temperature departures from assumed energy equilibrium (in 1765), with and without the RCP6 volcanic radiative forcings included.

The results show that 41% of the ocean warming in the model was simply due to the two major volcanoes early in the record. This is in good agreement with the 38% estimate from the Christy & McNider study.

It is interesting to see the “true” warming effects of the 1982-83 and 1991-1993 El Nino episodes, which were masked by the eruptions. The peak model temperatures in those events were only 0.1 C below the record-setting 1997-98 El Nino, and 0.2 C below the 2015-16 El Nino.

This is not a new issue, of course, as Christy & McNider also published a similar analysis in Nature in 1994.

These volcanic effects on the post-1979 warming trend should always be kept in mind when discussing the post-1979 temperature trends.

NOTE: In a previous version of this post I suggested that the Christy & McNider (1994) paper had been scrubbed from Google. It turns out that Google could not find it if the authors’ middle initials were included (but DuckDuckGo had no problem finding it).